A way to interconnect the multispectral images for improved vision systems by using the methods of diffusive morphology

Abstract

A new way to interconnect multispectral images by using the methods of diffusive morphology are presented for implementing improved vision systems in aviation. In contrast to the traditional morphological methods, in the presented approach, it is not necessary to segment in advance the image for describing its shape and it ensures strong resistance against noise and distortions. Procedures for two- and three-channel image interconnection are developed based on diffusive morphological filtration. How to use them by using the image pyramid is also presented. A promising prototype of an improved vision system is described. The interconnecting algorithms are processed experimentally by using the multispectral data collected during an aviation experiment by using the described prototype.